Tick Species Composition, Collection Rates, and Phenology Provide Insights into Tick-Borne Disease Ecology in Virginia

Abstract

To better understand tick ecology in Virginia and the increasing Lyme disease incidence in western Virginia, a comparative phenological study was conducted in which monthly collections were performed at twelve sampling locations in southwestern Virginia (high Lyme disease incidence) and 18 equivalent sampling locations in southeastern Virginia (low Lyme disease incidence) for one year. In western Virginia, we also explored the effect of elevation on collection rates of Ixodes scapularis Say (Acari: Ixodidae) and Amblyomma americanum (L.) (Acari: Ixodidae). In total, 35,438 ticks were collected (33,106 A. americanum; 2,052 I. scapularis; 134 Ixodes affinis Neumann [Acari: Ixodidae]; 84 Dermacentor variabilis [Say] [Acari: Ixodidae]; 49 Dermacentor albipictus [Packard] [Acari: Ixodidae]; 10 Haemaphysalis leporispalustris [Packard] [Acari: Ixodidae]; 2 Ixodes brunneus Koch [Acari: Ixodidae]; 1 Haemaphysalis longicornis Neumann [Acari: Ixodidae]). Within southwestern Virginia, Ixodes scapularis collection rates were not influenced by elevation, unlike A. americanum which were collected more frequently at lower elevations (e.g., below 500 m). Notably, I. scapularis larvae and nymphs were commonly collected in southwestern Virginia (indicating that they were questing on or above the leaf litter) but not in southeastern Virginia. Questing on or above the leaf litter is primarily associated with northern populations of I. scapularis. These findings may support the hypothesis that I. scapularis from the northeastern United States are migrating into western Virginia and contributing to the higher incidence of Lyme disease in this region. This comparative phenological study underscores the value of these types of studies and the need for additional research to further understand the rapidly changing tick-borne disease dynamics in Virginia.

Within the eastern United States, the four most common human-biting tick species are Ixodes scapularis (blacklegged tick), Amblyomma americanum (lone star tick), Dermacentor variabilis (American dog tick), and Amblyomma maculatum Koch (Acari: Ixodidae) (Gulf Coast tick) (Felz et al. 1996, Merten and Durden 2000, Stromdahl et al. 2001 , Goddard 2002 , Gleim et al. 2016). These tick species play key roles as vectors for pathogens that can cause diseases in humans and domestic animals (Oliver 1989). For example, I. scapularis is the primary vector for numerous human tick-borne pathogens, including Borrelia burgdorferi sensu stricto Johnson, Schmid, Hyde, Steigerwalt & Brenner (Spirochaetales: Borreliaceae) (Lyme disease), Borrelia miyamotoi Fukunaga, Takahashi, Tsuruta, Matushita, Ralph, McClelland, & Nakao (Spirochaetales: Borreliaceae) (Borrelia miyamotoi infection), Anaplasma phagocytophilum (Foggie) (Rickettsiales: Anaplasmataceae) (anaplasmosis), Babesia microti (Franca) (Piroplasmida: Babesiidae) (babesiosis), Ehrlichia muris eauclairensis Pritt, Allerdice, Sloan, Paddock, Munderloh, Rikihisa, Tajima, Paskewitz, Neitzel, Hoang Johnson, Schiffman, Davis, Goldsmith, Nelson, & Karpathy (Rickettsiales: Ehrlichiaceae) (ehrlichiosis), and Powassan virus (Amarillovirales: Flaviviridae) (Ogden et al. 2007 , Stromdahl and Hickling 2012, Sakamoto et al. 2014, Kocan et al. 2015). Meanwhile, A. americanum serves as a vector for Ehrlichia chaffeensis Anderson, Dawson, Jones, & Wilson (Rickettsiales: Anaplasmataceae), Ehrlichia ewingii Anderson, Greene, Jones, & Dawson (Rickettsiales: Anaplasmataceae), Panola Mountain Ehrlichia (all causing ehrlichiosis), Francisella tularensis (McCoy & Chapin) (Thiotrichales: Francisellaceae) (tularemia), Heartland virus (Bunyavirales: Phenuiviridae), and Bourbon virus (Articulavirales: Orthomyxoviridae) (Paddock and Yabsley 2007, Reeves et al. 2008, Brault et al. 2018, Godsey et al. 2021). Amblyomma americanum is also associated with Southern tick-associated rash illness (STARI) and alpha-gal allergy (Childs and Paddock 2003, Paddock and Yabsley 2007, Stromdahl and Hickling 2012, Commins and Pitts-Mill 2013, Steinke et al. 2015). Dermacentor variabilis harbors and transmits the bacteria F. tularensis and Rickettsia rickettsii (Wolbach) (Rickettsiales: Rickettsiaceae) (Rocky Mountain spotted fever; Burgdorfer 1975, Petersen et al. 2009, Stromdahl and Hickling 2012, Schumacher et al. 2015) and finally, A. maculatum is responsible for the transmission of Rickettsia parkeri Lackman, Bell, Stoenner, & Pickens (Rickettsiales: Rickettsiaceae), causing sporadic cases of spotted fever rickettsiosis (Lee et al. 2019).

To better understand and prevent tick-borne diseases in humans, it is crucial to know local tick species composition, abundance, and phenology. However, these variables can vary among geographic regions due to differences in biotic and abiotic factors such as host communities, elevation, weather, climate, and forest composition (Paddock and Yabsley 2007, LoGiudice et al. 2008, Allan et al. 2010, Cohen et al. 2010). Thus, by understanding resident tick species composition and activity within a particular region, more effective preventative measures can be developed to reduce the risk of tick-borne diseases in humans and domestic animals. Furthermore, this information can provide insights into shifting disease dynamics in a particular area, as well as variations in disease dynamics among different regions.

In the United States, over 35,000 confirmed and probable cases of Lyme disease are reported annually to the Centers for Disease Control and Prevention, although it is estimated that the actual number of cases may be closer to 476,000 (Center for Disease Control and Prevention 2021, Kugeler et al. 2021). This makes Lyme disease the most common tick-borne disease in the United States. Despite the established presence of I. scapularis populations throughout the eastern United States, Lyme disease has been historically concentrated in the Northeast, upper Midwest, and mid-Atlantic regions, with few autochthonous cases in the Southeast (Schwartz et al. 2017). However, over approximately the last fifteen years, Lyme disease cases have geographically expanded southward (Schwartz et al. 2017), including into Virginia (Lantos et al. 2015).

Specifically, since 2007, Lyme disease cases in Virginia have increased dramatically, with fewer than 400 cases per year between 2000 and 2006 but increasing to 957 cases in 2007 (Lantos et al. 2015). By 2016, the five-year running average was 1,265 Lyme disease cases annually in Virginia (Virginia Department of Health 2016). Interestingly, where these cases occur in the state has also shifted. Before 2007, the majority of Lyme disease cases occurred in the eastern portion of the state. However, since 2007, Lyme disease cases are now concentrating along the western regions of the state along the Appalachian Mountains (Brinkerhoff et al. 2014, Lantos et al. 2015). Counties in these high-elevation regions have shown the largest increase of Lyme disease incidence across the state, with some of the highest case numbers forming a distinct hotspot in southwestern Virginia (Brinkerhoff et al. 2014, Lantos et al. 2015, Seukep et al. 2015).

Therefore, it is important to better study tick species abundance, distribution, and phenology throughout Virginia, particularly western Virginia, to understand the rapid changes in Lyme disease dynamics within the state. In one study which examined I. scapularis nymphs in an east-west gradient across Virginia, I. scapularis nymphs were found at the highest densities and with the highest prevalence of B. burgdorferi sensu stricto at a single, high-altitude site in Nelson County located in western Virginia (Brinkerhoff et al. 2014). This, in conjunction with Lyme disease cases concentrating along the Appalachian Mountains, has led to the suggestion that I. scapularis is more abundant in western Virginia and may be generally more abundant at higher elevations. However, more extensive sampling in western Virginia is needed as one follow-up study conducted in the state did not find a significant relationship between I. scapularis abundance and elevation (Ferrell and Brinkerhoff 2018).

Importantly, no comprehensive phenological studies of ticks in Virginia have been published. These data are particularly critical in the southwestern region of the state, where the newly emerging Lyme disease hotspot is located. Furthermore, general data on all human-biting tick species are valuable as they provide insights on overall human tick-borne disease risk and could shed light on differences and/or shifts in tick ecology beyond Lyme disease. Therefore, the current study was undertaken in southeastern Virginia (low Lyme disease incidence) and at varying elevations in southwestern Virginia (newly emerging Lyme disease hotspot) to collect ticks monthly for one year. Taking a comparative approach, this study allowed us to: 1) better understand the driving forces behind the shift in Lyme disease dynamics within Virginia; 2) compare differences in tick species composition, relative collection rates, and phenology between two different geo-regions of the state (mountains versus coast); and 3) examine differences in tick collection rates (specifically I. scapularis and A. americanum) in relation to elevation.

Materials and Methods

Study Areas

Sample locations were established in mature forests typical of their respective regions in southwestern and southeastern Virginia (Fig. 1). Permits and permission from each respective agency (City of Roanoke, The Nature Conservancy, Roanoke County, Virginia Department of Wildlife Resources) and private landowners were received for all work done in the field. In southwestern Virginia, tick collections occurred at twelve sample locations in the City of Roanoke, and in Roanoke, Montgomery, Botetourt, and Bedford Counties (Fig. 1). These sample locations were deciduous and mixed pine-hardwood forests ranging from 340 to 570 m in elevation with oak, maple, ash, and pine species being the most common canopy-dominant species. Notably, the Appalachian Mountains are within this study region which also allowed us to examine the effects of elevation on various tick species.

Fig. 1.

Sampling locations for the southwestern and southeastern tick collections in Virginia, United States. Not all locations are visible as the symbols overlap.

The sample locations chosen for comparison in southeastern Virginia were ones that have been sampled at least monthly since 2009 as part of a long-term tick surveillance project (Nadolny et al. 2011, 2014; Wright et al. 2011, 2014a; Nadolny and Gaff, 2018; Espada et al. 2021) (Fig. 1). Data for this paper were limited to 18 sample locations ranging from 0-10 m in elevation in forested habitats typical of the region. Specifically, these forests were also primarily deciduous and mixed pine with some pine-dominated forests. The most common canopy-dominant species in the southeastern Virginia sites included oak and pine species.

Tick Collections

From February 2018 to January 2019, southwestern Virginia locations were sampled once a month, while in southeastern Virginia, locations were sampled once a month during the winter (when ticks were less active) and every two weeks during the rest of the year. All tick collections were performed using 1 m² flags. Tick collections were performed during daylight hours, only when the leaf litter and understory plants were dry, and collections were not performed during inclement weather conditions (e.g., rain, snow, excessive wind). Ticks were not collected in southeastern Virginia if the ambient temperature was below 5°C. Because of limited time for fieldwork and the cold temperatures experienced during winter in southwestern Virginia, there was no minimum ambient temperature cut-off for collecting ticks in this region. In southwestern Virginia, flagging was conducted in designated areas ranging from 0.2 to 0.4 ha which were marked with stakes and flagged in such a way that any given area within the location was not covered more than one time. Collection effort was standardized by time with flagging sessions lasting one man-hour. Flags were monitored constantly during flagging efforts and anytime a tick was observed on the flag, it was immediately collected. Furthermore, the flags were inspected regardless of whether a tick was perceived to be present every 5-min to ensure that no ticks were missed during the constant monitoring. Any time in which flagging was stopped for tick removal or to check the flag, the timer was stopped. All adult ticks and nymphs were removed from the flag and preserved in 70% ethanol. Larval ticks were collected en masse from the flag, using masking tape, and labeled by clutch. In southeastern Virginia, within each location, one fixed-distance transect (100–2,500 m; transect distance was dictated based on the size of the forest stand) was established in each spatially distinct forested stand, resulting in one to four transects per location. Flags were monitored constantly during flagging with ticks being removed as they were seen. Regardless of the perception of the presence or absence of ticks, the flag was also checked every 10-m. To convert the data to estimated collection rates per man-hour rather than distance, collections were retrospectively timed at each location multiple times by different collectors with the timer being stopped when ticks were being removed. Because there were minimal differences between times and collectors at each respective site, the average time per transect was then used to convert collection rates to time (e.g., per man-hour). Ultimately, the time spent flagging each transect over the span of a month varied from 10 min to just over one hour, depending on the length of the transect and the number of visits per month.

Tick Identification

Adults from all species and A. americanum nymphs were identified using published keys (Clifford et al. 1961, Keirans et al. 1989, Keirans and Durden 1998, Nadolny et al. 2021). Haemaphysalis spp. and Ixodes spp. nymphs and all larvae were identified to genus under compound light microscopes based on morphological characteristics (Clifford et al. 1961, Keirans et al. 1989, Keirans and Durden 1998). For southeastern sites, all Haemaphysalis spp. and Ixodes spp. nymphs and larvae were identified to species presumptively based on phenology and results of identification of species via molecular methods from previous years at those sites (Wright et al. 2014b). All of the southwestern Ixodes spp. and Haemaphysalis spp. nymphs, and a subset of the larvae (including all Haemaphysalis and Dermacentor spp.) were identified to species via a polymerase chain reaction (PCR) targeting the 16S mitochondrial rRNA gene and sequencing (Black and Piesman 1994). To do this, DNA extractions were performed using Qiagen DNeasy Blood and Tissue Extraction Kits (Qiagen, Valencia, CA) as per the manufacturer’s instructions. DNA was extracted from each individual Ixodes spp. nymph. For larvae, individuals from the same clutch were gathered in pools of 20 for extraction. Due to time and cost constraints, no more than 5 larval pools were extracted from any given site for each month. For sites that had more than 5 larval clutches in a given month, at least one pool of each genus was tested, and the remaining larvae were identified to species based on morphological observations in conjunction with sequencing results of other clutches from that date/ site.

The PCR was then performed using 16S+1 and 16S-2 primers targeting the mitochondrial 16S rRNA gene as in Black and Piesman (1994). Negative controls were used in all DNA extractions and PCR reactions. Both DNA extractions and PCR were performed in separate, designated biosafety cabinets in different rooms. Polymerase chain reaction products were then purified in a separate room according to manufacturer’s instructions using QIAquick Gel Extraction kits (Qiagen) and sent to Eurofin Genomics (Louisville, KY, USA) for sequencing.

Elevation and Vegetation Surveys

A previous field study conducted in 2004–2006 throughout the entire known range of I. scapularis proposed that I. scapularis did not occur above approximately 500-m elevation (Diuk-Wasser et al. 2010). However, the concentration of Lyme disease cases in western Virginia along the Appalachian Mountains which frequently have elevations above 500 m (Brinkerhoff et al. 2014, Lantos et al. 2015) necessitated further investigation into the impacts of elevation and other variables (e.g., forest structure) on I. scapularis abundance. To eliminate confounding variables associated with the distance between southeastern and southwestern regions (e.g., climate, differences in host-seeking behavior, etc.), and because there was minimal variation in elevation at the southeastern sites due to being located on the coastal plain, this aspect of the study was conducted only in southwestern Virginia. To do this, elevation was measured at each site in southwestern Virginia using a Pretel Alti Plus K2 altimeter (Alti-2, DeLand, FL) and further confirmed by examining our sites in Virginia digital elevation models (DEMs) (United States Geological Service 2017) in ArcMap 10.7 (Esri, Redlands, CA, USA). Seven sites were located below 500-m in elevation (340–488 m) and five sites were located above 500-m in elevation (507–570 m).

To consider other variables that may further impact I. scapularis abundance beyond elevation, vegetation surveys were also conducted at southwestern sites between June and August in 2018. Specifically, two to three (depending on the size of the site) 10 × 50 m belt transects were established at each site. Within each transect, canopy cover was taken using a spherical densitometer at four different locations within the transect and averaged. Also, within each transect, the diameter at breast height (DBH) of all trees >2.5 cm was measured. From this, DBH was converted to inches and total basal area was calculated using the following formula:

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{B}\mathrm{a}\mathrm{s}\mathrm{a}\mathrm{l}\mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a}=0.005454\mathrm{x}\mathrm{D}\mathrm{B}{\mathrm{H}}^2}\end{array}}$$

Statistical Analyses

Two core comparisons were examined: 1) I. scapularis and A. americanum counts by life stage in southwestern versus southeastern Virginia, and 2) the effects of elevation, basal area per acre (square-feet per acre), trees per acre, and canopy cover (%) on I. scapularis and A. americanum counts in southwestern Virginia. For all analyses, generalized estimating equations (GEE) negative binomial models were used to estimate the relative rates of the number of each life stage of each species collected per hour. Models included the number of ticks collected as the dependent variable and the natural logarithm of sampling time as a predictor with the coefficient constrained to one (i.e., as an offset variable). This particular approach allowed us to account for 1) repeated measures, 2) clustering by sites, 3) varying levels of collection efforts at our sites, and 4) the high number of zeros in our data set. For each tick species and analysis, models were created for each life stage and for total ticks. Because larvae numbers are so variable, the number of larval clutches was instead used as a measure for larval abundance in the model where a clutch was considered to be any larvae that were in close proximity to one another on the flag. No other tick species were collected in high enough numbers to model.

For models evaluating the effects of elevation, basal area, trees per acre, and canopy cover in southwestern Virginia, elevation was considered the primary predictor of interest. Other predictors were eliminated from the multivariable models based on their level of significance using a manual stepwise procedure until only variables having P < 0.05 remained. If the removal of a predictor resulted in greater than 20% change to the estimated effect of elevation, the predictor was considered a confounder and was retained in the model regardless of its significance. Season was modeled as a dichotomous predictor comparing warm (April—September) and cool (October—March) months. Two-way interactions between the predictor variables were not evaluated due to the sparsity of ticks at many sites during the cool season. All tests assumed a two-sided alternative hypothesis, and values of P < 0.05 were considered significant. Analyses were performed using commercially available statistical software (Stata version 16.1, StataCorp LLC, College Station, TX).

Results

Phenology

In southwestern Virginia, 19,434 ticks were collected from the field: 17,381 A. americanum (15,875 larvae, 1,419 nymphs, 87 adults), 1,936 I. scapularis (1,496 larvae, 306 nymphs, 134 adults), 59 D. variabilis (9 larvae, 50 adults), 49 Dermacentor albipictus (larvae), 6 Haemaphysalis leporispalustris (5 larvae, 1 nymph), 2 Ixodes brunneus (larvae), and 1 Haemaphysalis longicornis (larva) (Table 1). In southeastern Virginia, a total of 16,004 ticks were collected: 15,725 A. americanum (14,818 larvae, 637 nymphs, 270 adults,), 134 Ixodes affinis (adults), 116 I. scapularis (54 larvae, 13 nymphs, 49 adults), 25 D. variabilis (adults), and 4 H. leporispalustris (nymphs) (Table 1). Seasonal trends were determined for all life stages of I. scapularis and A. americanum in southwestern Virginia and southeastern Virginia during the, February 2018 to January 2019 study period (Fig. 2). In southwestern Virginia, I. scapularis adult activity exhibited a weak bimodal peak where activity began in October, peaked in November and December, and declined in January. A second round of activity commenced and peaked in February and ended in April. Ixodes scapularis adult activity in southeastern Virginia also began in October, but peaked November through January and continued for a slightly longer period of time, extending into May (Fig. 2). For I. scapularis nymphs, there was a longer activity period in southwestern Virginia, with nymphs being collected from April through October with activity peaking May through July. In southeastern Virginia, nymphs were only active from May through August, with activity peaking in June. Ixodes scapularis larvae exhibited a bimodal peak in southeastern Virginia, with activity occurring from April to June with peak activity being May through June (and a smaller amount of activity in August). Conversely, I. scapularis larvae in southwestern Virginia had a single activity period between the months of June and October with peak activity occurring July through September (Fig. 2).

Table 1.

Tick species and life stages collected from February 2018 to January 2019 in southwestern Virginia (SW VA) and southeastern Virginia (SE VA), United States. Hours of collection in southwestern Virginia = 144, hours in southeastern Virginia = 57

Species	Life stage	Months Collected in SW VA	Months Collected in SE VA	Total # in SW VA (Ticks/hr)	Total # in SE VA (Ticks/hr)
Ixodes scapularis	Adult	Jan—Mar, Oct—Dec	January—May, October—December	134 (0.9)	49 (0.9)
Ixodes scapularis	Nymph	Apr—Oct	May—August	306 (2.1)	13 (0.2)
Ixodes scapularis	Larvae	June—Oct	June—November	1,496 (10.4)	54 (1.0)
Ixodes brunneus	Larvae	Jan	---	2 (0.01)	0
Ixodes affinis	Adult	–	April—August, October	0	134 (2.4)
Ambylomma americanum	Adult	Feb—July	April—August	87 (0.6)	270 (4.8)
Ambylomma americanum	Nymph	Apr—Oct	April—October	1,419 (9.8)	637 (11.2)
Ambylomma americanum	Larvae	May—Nov	May, July—November	15,875 (110.2)	14,818 (261.6)
Dermacentor variabilis	Adult	April—Aug	April—July	50 (0.4)	25 (0.4)
Dermacentor variabilis	Larvae	Aug	–	9 (0.06)	0
Dermacentor albipictus	Larvae	Oct—Nov	–	49 (0.3)	0
Haemaphysalis leporispalustris	Nymph	June	March, May, October	1 (0.01)	4 (0.07)
Haemaphysalis leporispalustris	Larvae	Sept—Nov	–	5 (0.03)	0
Haemaphysalis longicornis	Larvae	Oct	–	1 (0.01)	0
Total				19,434	16,936

Fig. 2.

Mean number of Ixodes scapularis and Amblyomma americanum ticks collected per hour by life stage and month in southwestern Virginia (n= 12) and southeastern Virginia (n = 18), United States. Note, one clutch of larvae counted as one tick. Panels summarize results by location and tick species.

Similar seasonal trends were observed between southwestern and southeastern Virginia for all life stages of A. americanum during the study period with a few minor differences (Fig. 2). Specifically, A. americanum adults were active in both regions from April to July. However, southwestern Virginia also had minor activity in February and a peak in activity from May through July. In southeastern Virginia, the peak activity period for A. americanum adults occurred in May. In both regions, A. americanum nymphs were active from April to October. Again, southwestern Virginia had a longer period of peak nymphal activity (May through July), whereas the nymphal activity for A. americanum in southeastern Virginia peaked in May. Most A. americanum larval activity in both southwestern and southeastern Virginia occurred between May to October and peaked in August and September (Fig. 2). However, activity in southwestern Virginia occurred for a somewhat longer period of time with slight activity beginning in May and ending with minor activity in November.

During the study period, in southwestern Virginia, adult D. variabilis activity began in April, peaked in May, and activity continued through the end of August. Similar trends were observed in southeastern Virginia, where adult D. variabilis activity peaked in May and ended in July. Due to their low numbers, seasonal trends could not be determined for D. variabilis larvae, D. albipictus, I. brunneus, H. leporispalustris, and H. longicornis (Table 1).

Collection Rates

Collection rates between life stages of both I. scapularis and A. americanum varied both seasonally and between southwestern Virginia (e.g., high Lyme disease incidence) and southeastern Virginia (e.g., low Lyme disease incidence). Ixodes scapularis nymphal and larval collection rates were significantly higher in southwestern Virginia than southeastern Virginia, whereas there was no significant difference in I. scapularis adult collection rates between these regions (Table 2, Fig. 3). Amblyomma americanum adult collection rates were significantly higher in southeastern Virginia than southwestern Virginia. However, A. americanum nymphs and larvae collection rates were not significantly different between the two regions (Table 3, Fig. 3).

Table 2.

Generalized estimating equation negative binomial regression model for the prediction of collection rates for life stages of Ixodes scapularis between southwestern Virginia (SW VA) and southeastern Virginia (SE VA), USA, from February 2018 to January 2019

Tick Species, life stage	Variable	Coefficient (SE)	RR (95% CI)	P
Ixodes scapularis, All life stages	Season (warm vs. coola)	0.77 (0.34)	2.17 (1.12, 4.21)	0.022
	Location (SW VA vs. SE VAa)	1.31 (0.45)	3.69 (1.54, 8.88)	0.004
	Constant	−0.31 (0.29)	NA	0.293
Ixodes scapularis, Adult	Season (warm vs. coola)	−2.50 (0.33)	0.08 (0.04, 0.16)	<0.001
	Location (SW VA vs. SE VAa)	−0.44 (0.73)	0.64 (0.15, 2.66)	0.541
	Constant	1.01 (0.50)	NA	0.046
Ixodes scapularis, Nymph	Season (warm vs. coola)	3.63 (0.55)	37.68 (12.81, 110.85)	<0.001
	Location (SW VA vs. SE VAa)	2.41 (0.61)	11.15 (3.35, 37.04)	<0.001
	Constant	−4.62 (0.68)	NA	<0.001
Ixodes scapularis, Larvae	Season (warm vs. coola)	4.30 (0.98)	73.60 (10.77, 502.78)	<0.001
	Location (SW VA vs. SE VAa)	0.82 (0.27)	2.28 (1.33, 3.89)	0.003
	Constant	−4.09 (1.01)	NA	<0.001

SE = standard error. RR = relative rate. CI = confidence interval. NA = not applicable.

^a indicates reference category.

Fig. 3.

Marginal predicted number (±SE) of Ixodes scapularis and Amblyomma americanum ticks collected per hour by species, life stage, and location (averaged across all 12 mo) for 12 sites in southwestern Virginia and 18 sites in southeastern Virginia, United States, based on generalized estimating equations negative binomial regression models. Note, one clutch of larvae counted as one tick ^***P < 0.001, ^**P < 0.01.

Table 3.

Generalized estimating equation negative binomial regression model for the prediction of collection rates for life stages of Amblyomma americanum between southwestern Virginia (SW VA) and southeastern Virginia (SE VA), USA, from February 2018 to January 2019

Tick Species, life stage	Variable	Coefficient (SE)	RR (95% CI)	P
Amblyomma americanum, All life stages	Season (warm vs. coola)	3.72 (0.31)	41.19 (22.42, 75.67)	<0.001
	Location (SW VA vs. SE VAa)	−0.26 (0.52)	0.77 (0.28, 2.13)	0.621
	Constant	−0.31 (0.38)	NA	0.417
Amblyomma americanum, Adult	Season (warm vs. coola)	3.18 (0.56)	24.12 (8.06, 72.16)	<0.001
	Location (SW VA vs. SE VAa)	−1.90 (0.49)	0.15 (0.57, 0.39)	<0.001
	Constant	−1.23 (0.58)	NA	0.033
Amblyomma americanum, Nymph	Season (warm vs. coola)	5.61 (0.56)	272.38 (90.43, 820.38)	<0.001
	Location (SW VA vs. SE VAa)	0.91 (0.57)	1.10 (0.35, 3.38)	0.874
	Constant	−2.70 (0.64)	NA	<0.001
Amblyomma americanum, Larvae	Season (warm vs. coola)	1.89 (0.33)	6.65 (3.45, 12.82)	<0.001
	Location (SW VA vs. SE VAa)	−0.40 (0.40)	0.67 (0.30, 1.46)	0.313
	Constant	−0.53 (0.37)	NA	0.158

SE = standard error. RR = relative rate. CI = confidence interval. NA = not applicable.

^a indicates reference category.

Elevation and Forest Characteristics

Collection rates of I. scapularis adults, nymphs, and larvae did not vary significantly with elevation (Table 4, Fig. 4). Additionally, I. scapularis larvae showed a statistically significant negative association with trees per acre, indicating that this life stage was either less abundant and/or less likely to be questing on or above the leaf litter as tree density increased (Table 4). Interestingly, all life stages of A. americanum exhibited a negative association with elevation (Fig. 4), and A. americanum nymphs and larvae displayed a negative association with canopy cover (Table 5). Amblyomma americanum nymphs also exhibited a statistically significant positive association with basal area within study plots. Additionally, as was observed with I. scapularis larvae, A. americanum larvae also exhibited a negative association with tree density.

Table 4.

Generalized estimating equation negative binomial regression model for the prediction of collection rates for life stages of Ixodes scapularis in relation to elevation and forest characteristics in southwestern Virginia, USA, from February 2018 to January 2019

Tick Species, life stage	Variable	Coefficient (SE)	RR (95% CI)	P
Ixodes scapularis, All life stages	Season (warm vs. coola)	1.31 (0.23)	3.69 (2.38, 5.75)	<0.001
	Elevation (m)	−0.06 (0.19)	0.94 (0.65, 1.35)	0.736
	Constant	0.92 (1.15)	NA	0.419
Ixodes scapularis, Adult	Season (warm vs. coola)	−2.68 (0.40)	0.07 (0.03, 0.15)	<0.001
	Elevation (m)	0.56 (0.40)	1.76 (0.80, 3.86)	0.160
	Trees per acre	−0.55 (0.40)	0.57 (0.26,1.26)	0.166
	Constant	0.47 (2.01)	NA	0.815
Ixodes scapularis, Nymph	Season (warm vs. coola)	3.61 (0.58)	37.23 (12.04, 115.13)	<0.001
	Elevation (m)	0.04 (0.26)	1.04 (0.63, 1.73)	0.881
	Constant	−2.37 (1.26)	NA	0.060
Ixodes scapularis, Larvae	Season (warm vs. coola)	4.21 (1.02)	67.37 (9.04, 501.84)	<0.001
	Elevation (m)	−0.07 (0.21)	0.93 (0.62, 1.40)	0.740
	Trees per acre	−0.31 (0.07)	0.74 (0.64, 0.84)	<0.001
	Constant	−1.35 (1.57)	NA	0.389

SE = standard error. RR = relative rate. CI = confidence interval. NA = not applicable.

^a indicates reference category.

Fig. 4.

Scatter plots (and linear fit) of elevation versus the annual mean number of Ixodes scapularis and Amblyomma americanum ticks collected per hour for 12 sites in southwestern Virginia, United States. Panels summarize results by tick species and life stage. Note, one clutch of larvae counted as one tick.

Table 5.

Generalized estimating equation negative binomial regression model for the prediction of collection rates for life stages of Amblyomma americanum in relation to elevation and forest characteristics in southwestern Virginia, USA, from February 2018 to January 2019

Tick Species, life stage	Variable	Coefficient (SE)	RR (95% CI)	P
Amblyomma americanum, All life stages	Season (warm vs. coola)	3.52 (0.34)	33.66 (17.36, 65.26)	<0.001
	Elevation (m)	−0.85 (0.38)	0.43 (0.20, 0.89)	0.024
	Canopy (%)	−0.28 (0.11)	0.75 (0.60, 0.94)	0.013
	Basal Area (sq ft per ac)	0.04 (0.01)	1.04 (1.02, 1.06)	0.001
	Constant	22.97 (10.54)	NA	0.029
Amblyomma americanum, Adult	Season (warm vs. coola)	2.32 (0.46)	10.13 (4.13, 24.83)	<0.001
	Elevation (m)	−0.78 (0.35)	0.46 (0.23, 0.91)	0.026
	Trees per acre	−0.60 (0.14)	0.55 (0.42, 0.72)	<0.001
	Constant	3.85 (1.72)	NA	0.025
Amblyomma americanum, Nymph	Season (warm vs. coola)	5.86 (0.68)	351.37 (92.90, 1329.04)	<0.001
	Elevation (m)	−0.92 (0.41)	0.40 (0.18, 0.88)	0.024
	Canopy (%)	−0.28 (0.14)	0.76 (0.58, 0.99)	0.045
	Basal Area (sq ft per ac)	0.04 (0.01)	1.04 (1.02, 1.07)	<0.001
	Constant	19.08 (12.33)	NA	0.122
Amblyomma americanum, Larvae	Season (warm vs. coola)	1.62 (0.50)	5.04 (1.90, 13.34)	0.001
	Elevation (m)	−0.78 (0.32)	0.46 (0.25, 0.86)	0.015
	Canopy (%)	−0.29 (0.12)	0.75 (0.60, 0.81)	0.012
	Trees per acre	−0.34 (0.08)	0.70 (0.60, 0.81)	<0.001
	Constant	29.96 (10.92)	NA	0.006

SE = standard error. RR = relative rate. CI = confidence interval. NA = not applicable.

^a indicates reference category.

Discussion

To our knowledge, this is the first study performed in southwestern Virginia on tick phenology and is the first comparative study of tick ecology in southeastern versus southwestern Virginia. While phenology and collection rates will change somewhat from year to year and additional years of data are needed to further confirm these trends, this still provided insights on tick species composition, relative collection rates, and potential differences in phenology within southeastern and southwestern Virginia. In addition to I. scapularis and A. americanum, other tick species collected during this study included D. variabilis, D. albipictus, I. brunneus, H. leporispalustris, and H. longicornis, which have all been previously reported in the state (Bishopp and Trembley 1945, Sonenshine and Stout 1970, Beard et al. 2018). While H. longicornis has been reported in other western Virginia counties, it had not been reported in any of the counties of our southwestern study sites (Beard et al. 2018). Thus, the single H. longicornis larva collected in this study was the first official report of this invasive species within Botetourt County, Virginia. Continued monitoring of this species’ geographical expansion and increasing abundance in the state of Virginia will be critical due to both agricultural and human health concerns.

There were no significant correlations with the collection rates of any life stage of host-seeking I. scapularis and elevation in southwestern Virginia, thus refuting past speculation that I. scapularis may be more abundant at high elevation sites in western Virginia (Brinkerhoff et al. 2014). However, the range of elevations examined in the southwestern sites was fairly small (340–570 m). Examining a wider range of elevations across western and central Virginia may reveal correlations not observed in the current study. The study did, however, confirm that I. scapularis populations are present above 510 m despite Diuk-Wasser et al. (2010) previously finding that I. scapularis did not occur above this altitude. The fact that I. scapularis was not collected above 510 m by Diuk-Wasser et al. (2010) may be due to having few sites above 510 m in the southern part of I. scapularis range. It’s also possible that they were simply sampling prior to the southern expansion of I. scapularis populations and that I. scapularis is better able to survive at higher elevations in Virginia due to the generally milder winters and/or a longer warm season at high elevations in Virginia as compared to more northern latitudes. Although less likely, it’s also possible that increased annual temperatures due to climate change over the past ten years since the Diuk-Wasser et al. (2010) study have allowed I. scapularis populations to expand into higher elevations within the state.

While I. scapularis did occur at high elevations, all life stages of A. americanum were collected at significantly lower rates at these elevations (e.g., over 500-m) which indicates a lower risk of pathogens vectored by A. americanum at these higher elevations. Notably, A. americanum adult collection rates were significantly higher in southeastern versus southwestern Virginia. This difference could be due to almost half of the collection sites in southwestern Virginia being located at high elevation sites where A. americanum populations were rarely collected, thus reducing the overall collection rates in this region. An alternative reason for the higher abundance of A. americanum in southeastern Virginia could be differences in host populations, climate and/or weather.

Microclimatic conditions at higher elevations are generally characterized by colder temperatures, higher winds, and lower relative humidity. Notably, A. americanum is known to cease water vapor intake below 5°C (Knulle and Rudolph 1982). Thus, A. americanum may not be able to maintain proper moisture levels in the colder temperatures associated with higher altitude environments. Further investigation of differences in the relative cold tolerance of I. scapularis and A. americanum, along with their physiological and behavioral drivers, may provide further insight into how each species adapts to distinct microclimatic conditions and better predict future range expansions of these species.

Numerous studies have found significant correlations with tick abundance and forest structure measurements (Alder et al. 1992, Ostfeld et al. 1995, Gleim et al. 2014, Mathisson et al. 2021). The current study found that A. americanum was associated with sites with lower tree density and canopy cover, and sites with higher basal area. Thus, these forest stand characteristics generally indicate that A. americanum is associated with mature forests. Notably, higher canopy cover would generally be associated with mature forests. However, canopy cover measurements were taken approximately 1-m above the ground. Because of this, higher canopy cover measurements were generally associated with younger forest stands which had a denser midstory which was captured via our canopy measurements. The fact that A. americanum were found in greater abundance in more mature forests may be due to these large, mature trees producing large amounts of nuts and acorns which are important food sources for wildlife hosts of A. americanum, like white-tailed deer (Odocoileus virginianus Zimmermann [Artiodactyla: Cervidae]) (Patterson and Power 2002, Anderson et al. 2003, Russell et al. 2017).

In other findings, there were some notable differences in the activity periods of I. scapularis nymphs and larvae in southwestern versus southeastern Virginia. Specifically, during the study period, nymphal activity occurred before larval activity in southwestern Virginia, whereas, in southeastern Virginia, nymphal and larval activity occurred at roughly the same time with the majority of larval activity preceding peak nymphal activity. While additional years of data are needed to determine whether this trend consistently occurs each year, these asynchronous activity periods observed in southwestern Virginia in which nymphs are active before larvae are more in-line with what occurs in the northeastern U.S. (Fish 1993, Gatewood et al. 2009, Ogden et al. 2018) whereas larval activity occurring synchronously, if not slightly before nymphal activity is more in-line with what has been reported in some southeastern states (Rogers 1953, Clark et al. 1998, Ogden et al. 2018). These differences in phenology have implications for enzootic pathogen transmission and disease risk in humans. Specifically, it’s been previously hypothesized that when nymphal activity occurs before larvae (as was observed in southwestern Virginia), nymphs are able to infect competent wildlife hosts which larvae later feed on and become infected themselves (Fish 1993). Not only does this allow for a sustained sylvatic cycle but furthermore, this asynchronous nymphal and larval activity has been shown to select for strains of B. burgdorferi that are associated more frequently with disseminated Lyme disease (Gatewood et al. 2009). This is in contrast to larvae being active before nymphal activity in southeastern Virginia. In this case, it’s been previously hypothesized that larvae would therefore be less likely to feed on an infected host and therefore halt any sort of sylvatic cycle (Oliver 1996). These differences in nymphal and larval phenology between the southwestern and southeastern portions of the state could potentially play an important role in the higher prevalence of B. burgdorferi sensu stricto documented by Brinkerhoff et al. (2014) in I. scapularis nymphs in western Virginia and be driving, at least in part, the larger number of human cases of Lyme disease in this part of the state. Pathogen testing of the ticks in western Virginia is planned to further confirm whether there are higher pathogen prevalences throughout the western region of Virginia where Lyme incidence is high.

In comparing I. scapularis mean ± SE collection rates between southwestern and southeastern Virginia, I. scapularis nymphs in southwestern Virginia have a high peak in activity (12 ± 6.2 nymphs/hr) that was not observed in southeastern Virginia (0.4 ± 0.2 nymphs/hr). Further confirming this trend of higher nymphal activity in western Virginia, Brinkerhoff et al. (2014) also found that their most western site in Virginia had the highest abundance of I. scapularis nymphs (9.6 nymphs/200 m²) as compared to sites in central and eastern Virginia which had fewer I. scapularis nymphs (1.7 nymphs/200 m², 1.1 nymphs/200 m², and 0.2 nymphs/200 m²). Additional fieldwork performed in central Virginia has further confirmed that I. scapularis nymphs are rarely collected in this region of the state (Jory Brinkerhoff, University of Richmond, personal communication). The scarcity of nymphs in central and eastern Virginia, in conjunction with similar numbers of I. scapularis adults between southwestern and southeastern Virginia, seems to indicate different host-seeking activity of nymphs (e.g., more likely to quest on or above the leaf litter) in southwestern Virginia rather than differences in abundance between the two locations.

Importantly, nymphal host-seeking behavior on or above leaf litter rather than within the leaf litter is associated with northern populations of I. scapularis (Arsnoe et al. 2015), whereas southern populations host-seek within the leaf litter where they are less likely to come into contact with humans or to be collected via flagging. Thus, it has been hypothesized that this difference in host-seeking behavior between northern and southern I. scapularis is at least one of the contributing factors to the disparity in Lyme disease incidence between the northeastern and southeastern United States (Arsnoe et al. 2019). Interestingly, based on genetic analyses, it has been hypothesized that I. scapularis populations in Virginia, particularly western Virginia, have migrated southwards from the northeastern United States (Brinkerhoff et al. 2014, Kelly et al. 2014). Although additional genetic analyses are needed to confirm this hypothesis, the northern host-seeking behavior exhibited by nymphs in southwestern Virginia in the current study seems to further support this hypothesis. If I. scapularis in western Virginia have migrated from the Northeast, it is also possible that they are better adapted to colder temperatures than southern populations and this may be another reason why they are better able to survive at higher elevations.

There have been minimal phenological studies performed that have reported on the activity of I. scapularis larvae and none within the state of Virginia. While transovarial transmission of B. burgdorferi does not occur, I. scapularis larvae can vector other pathogens such as Borrelia miyamotoi (Han et al. 2019). Therefore, the significantly higher collection rates of I. scapularis larvae in southwestern Virginia indicate higher risk in this region of the state for human infections associated with pathogens vectored by I. scapularis larvae such as B. miyamotoi.

The significantly higher I. scapularis peak larval activity in southwestern Virginia (90 ± 23 larvae/hr) in comparison to southeastern Virginia (2.5 ± 1.4 larvae/hr) may also have broader implications regarding Lyme disease ecology. Although previous studies have reported differences in the host-seeking behavior of northern and southern populations of I. scapularis nymph populations (Arsnoe et al. 2015, 2019), to our knowledge, there have been no published reports that have explicitly shown or discussed implications of the differences in host-seeking behavior of northern versus southern populations of I. scapularis larvae. However, past studies conducted across the eastern United States using environmental sampling methods (i.e., flagging, dragging, and CO₂ trapping) have shown that I. scapularis larvae are collected regularly in northern states (Stafford and Magnarelli 1993, Daniels et al. 1996, Solberg et al. 2002, Khatchikian et al. 2015), compared to little or no I. scapularis larvae collected in southern states (Casteel and Sonenshine 1996, MacKay and Foil 2005, Goddard and Piesman 2006, Gleim et al. 2014). Notably, Gleim et al. (2014) utilized identical collection methods to the current study and while similar numbers of I. scapularis adults were collected during their peak activity period in southwestern Georgia/ northwestern Florida versus southwestern Virginia (2.3 ± 1.5 larvae per hour versus 3.4 ± 1.6 adults per hour), vastly different numbers of I. scapularis larvae were collected during their peak activity period with an average of 0.2 ± 0.2 larva per hour in Georgia/Florida (only a single clutch of 5 individuals was collected over the course of the entire study) versus 90 ± 23 larvae/hour larvae per hour in southwestern Virginia. Again, the similar number of adult I. scapularis at both locations along with the very different numbers of I. scapularis larvae likely indicates differences in host-seeking behavior, not abundance. This increased activity of I. scapularis larvae in Southwest Virginia is notable for two reasons, 1) it seems to further support the hypothesis that northern populations of I. scapularis are migrating south along the Appalachian Mountains into Virginia in the western portion of the state, and 2) it has implications for differences in larval host selection which is likely playing a role in the differences in Lyme disease incidence between southeastern and southwestern Virginia and more broadly, between the northeastern and southeastern United States.

Numerous studies have indicated northern and southern populations of I. scapularis larvae differ greatly in the host species that they utilize (Kerr 2012; Goddard et al. 2015; Ginsberg et al. 2017, 2021). Specifically, northern populations of I. scapularis larvae are generally known to feed primarily on small mammals, particularly white-footed mice (Peromyscus leucopus) and shrews (Blarina spp.) which are competent reservoirs for B. burgdorferi sensu stricto (Donahue et al. 1987, Goodwin et al. 2001, LoGiudice et al. 2003, Brisson et al. 2008, Ginsberg et al. 2021). Conversely, southern populations of I. scapularis larvae typically feed on lizards which are not competent reservoirs for B. burgdorferi sensu stricto (Apperson et al. 1993, Kerr 2012, Ginsberg et al. 2021). Host-seeking behavior of larvae in southwestern Virginia is more similar to northern populations of I. scapularis (e.g., questing on or above the leaf litter). This likely indicates that I. scapularis larvae in southwestern Virginia will be more likely to utilize small mammal hosts (which are more competent reservoirs for B. burgdorferi sensu stricto). Should this occur, it would result in an increased prevalence of B. burgdorferi sensu stricto in nymphs. Conversely, larvae in southeastern Virginia may be more likely to utilize lizards (e.g., eastern fence lizard Sceloporus undulatus (Bosc & Daudin) (Squamata: Phrynosomatidae) and five-lined skinks Eumeces [Plestiodon] fasciatus Linnaeus [Squamata: Scincidae]) as hosts.

While it is still unclear what drives the differences observed in host species utilization between northern and southern I. scapularis populations, a recent study found that these differences are more likely due to distinct host preferences rather than differences in host species diversity and/or composition (Ginsberg et al. 2021). Further research into what drives host preferences is still needed. However, one study found that southern I. scapularis ticks showed no preferences for mice versus lizards in the lab (James and Oliver 1990), suggesting that host preference is not driven solely by chemical cues. We hypothesize that differences in host-seeking behaviors of larvae and nymphs observed in the current study, in conjunction with the potential differences in where these different host species spend their time (e.g., white-footed mice may spend more time above the leaf litter whereas skinks may spend more time within the leaf litter), could be contributing to the differences in host associations observed in I. scapularis in the northern versus southern U.S. Further research should be done to explore these hypotheses, as this may help further explain the higher incidence of Lyme disease in western Virginia as well as better understand the drivers between the differences in Lyme case incidence between the northern and southern U.S.

Collectively, these findings indicate that rapidly changing Lyme disease ecology in western Virginia is likely due to an ongoing southern expansion of northern populations of I. scapularis from the Northeast into the Southeast via the Appalachian Mountains. In addition to the northern host-seeking behavior being exhibited by I. scapularis nymphs and larvae and higher pathogen prevalence previously identified in I. scapularis nymphs (Brinkerhoff et al. 2014), there could be other factors contributing to the increasing Lyme disease incidence in southwestern Virginia. For example, the Roanoke Valley in southwestern Virginia where this study was conducted has some of the highest human population densities in western Virginia (Roanoke: 893 people/km², Blacksburg: 832 people/km²; US Census 2020). Given the presence of the Appalachian Mountains, many residents in this outdoor-centric region spend significant amounts of time in ideal tick habitat using the many outdoor recreational resources in the area (e.g., the Appalachian trail, national parks, or nationally acclaimed mountain biking trails). This prevalence of outdoor activity, in conjunction with growing human population sizes and I. scapularis nymphs more willing to ascend vegetation to find a host, may, in turn, translate to higher Lyme disease risk for residents of Southwest Virginia. Therefore, future studies should work to understand the activities undertaken by citizens residing in southwestern Virginia and how this impacts their risk for Lyme disease.

Conclusion

Overall, this work underscores the value of phenological studies, particularly comparative ones, as they help provide valuable insights into shifting disease dynamics and provide critical information regarding tick-borne disease risk and mitigation. These results 1) provide coarse information on activity periods and relative collection rates of tick species in these two regions of Virginia, and 2) indicated that I. scapularis larvae and nymphs are host-seeking on or above the leaf litter in Southwest Virginia which is not occurring in other parts of the state (Brinkerhoff et al. 2014, Jory Brinkerhoff, University of Richmond, personal communication). This host-seeking behavior, in addition to the large human population in southwestern Virginia which frequently recreates outdoors, is likely contributing to the unusually high Lyme disease incidence in this region. Furthermore, this study proposes the important role that the host-seeking behavior of I. scapularis larvae, in addition to nymphs, may play in differing dynamics of enzootic transmission of Lyme disease spirochetes in the northeastern and southeastern United States.

Ultimately, past studies on I. scapularis and B. burgdorferi dynamics in Virginia, North Carolina, and Tennessee have indicated that the frontline of high Lyme disease incidence is expanding southward (Brinkerhoff et al. 2014, Kelly et al. 2014, Lantos et al. 2015, Hickling et al. 2018). Our findings that I. scapularis nymphs and larvae in southwestern Virginia are exhibiting host-seeking behavior typically associated with northern I. scapularis seems to further support this assertion. However, other studies are needed to 1) compare phenological activity over multiple years in Virginia to more definitively explore and track differences between the eastern versus western part of Virginia, 2) understand whether there are additional drivers behind high Lyme disease incidence in western Virginia (e.g., high pathogen prevalence, wildlife host dynamics, and/or human behavior), and 3) further explore whether northern populations of I. scapularis are indeed migrating southwards into Virginia and if so, what is driving and facilitating this migration.